Filing in practice
The source documents contain no checklist of criteria for “having public opinion attributes or social mobilisation capacity”, and no quantitative threshold. This judgement is directly tied to the filing and security assessment obligations; the right approach is to make a written assessment against the dimensions listed in the provisions and confirm it with the local cyberspace administration department, rather than applying standards circulating online.
Article 23 of the algorithm recommendation provisions gives the dimensions of tiered and classified management: cyberspace administration departments, together with telecommunications, public security, market regulation and other relevant departments, shall establish a tiered and classified algorithm safety management system, and shall implement tiered and classified management of algorithm recommendation service providers according to the public opinion attributes or social mobilisation capacity of the algorithm recommendation service, the content category, the user scale, the importance of the data processed by the algorithm recommendation technology, the degree of intervention in user behaviour, and so on. These five dimensions are the only handle in the provisions for organising the basis of a judgement.
Both Article 24 and Article 27 use “public opinion attributes or social mobilisation capacity” as the trigger: the former triggers the filing formalities and the latter triggers the security assessment. Article 17 of Interim Measures for the Administration of Generative AI Services uses the same qualifier, triggering a security assessment and pointing to algorithm filing.
The approach can therefore be summarised in three steps: first, describe clearly how the service disseminates information (whether it disseminates information to unspecified members of the public, and whether it has organising and mobilising features); second, fill in the facts item by item against the dimensions listed in Article 23; third, submit that assessment conclusion to the local cyberspace administration department for confirmation. The source documents neither give judging criteria nor designate who judges, and this knowledge base makes no presumption.
Deducing “no public opinion attributes” straight from “the user base is small” is inventing your own standard. Reviewing only content-type products and ignoring the social mobilisation capacity that public sharing, rankings and comment areas can bring in tool-type products. Treating generative AI services and algorithm recommendation services as two unrelated sets of duties. Letting the conclusion exist only orally, with no assessment material that can be produced, so the reasoning cannot be explained at inspection.
As for how to complete the “content to be published” field in the filing system, the provisions only require the field to be filed and do not prescribe a format; the local cyberspace administration authority's practice prevails.
This entry belongs to the “China AI Compliance and Filing” cluster (42 entries in total): View all · Answers home
Who we are: a knowledge base and AI customer service team serving small and micro businesses, delivering implementations that take AI customer service through the national standard and the filing process.
Content on this site is compiled from publicly released regulatory texts for enterprises' own self-check reference; the specific filing requirements of the local cyberspace administration authorities prevail, and this is not legal advice.
This page is generated from the China AI compliance and filing MCP corpus — agents can call the same data directly.